一种针对LSB匹配隐写的负载定位新算法

针对最不重要比特位(Least significant bit, LSB)匹配隐写算法,本文提出了一种新的负载定位算法。将隐写负载定位看作二分类问题,将载密图像每个像素位置看作待分类样本,通过提取载密图像集中每个像素位置在8个方向上的相邻像素差分平方均值特征,利用支持向量机(Support vector machine,SVM)分类器,将每个像素位置划分到正确的类别——负载位置或非负载位置。本文从理论和实验两方面验证了所提分类特征的有效性。针对LSB匹配隐写,本文方法与最大后验概率(Maximum a posteriori, MAP)载体估计方法做出比较,在低嵌入率条件下,本文方法的定位性能有明显提高。     

基于LSBMR图像边缘自适应隐写的检测

在图像边缘自适应LSB匹配改进隐写算法中,秘密信息嵌入位置的选择仅由某个方向上像素对的差值决定,未考虑该像素与其邻域内其他像素的差值变化的特点.针对该问题,对隐写前后图像的八方向差分直方图进行分析,提出一种基于LSB匹配改进算法(LSBMR)边缘自适应隐写检测的算法.该算法计算图像的八方向绝对差分直方图,提取直方图中隐写前后变化较为明显的频数用以构建特征向量,并使用支持向量机完成检测.对较低嵌入率下(≤0.5 bpp)的EALSBMR隐写结果进行检测,结果表明该算法的平均检测率均高于现有典型的隐写分析算法.     

基于逐像素概率预测的图像隐写定位研究

为进一步增强图像隐写分析的实用性,对内容自适应隐写术和非内容自适应LSB matching的隐写像素定位问题展开研究,提出一种端到端的图像隐写定位网络PSL_NET,在输入端输入一张图像,输出端定位出图像的隐写像素的位置。在预处理层中,利用空域富模型的高通滤波器提取残噪图像;在深度残差层中,通过深度残差学习增强隐写特征的表达能力;在像素预测层中,利用标记出隐写像素实际位置的掩码图像进行有监督地学习,增强网络对局部隐写像素的感知能力,无区别对待平滑或者纹理区域的像素,逐一预测图像每位像素是真实位或是隐写位,最终预测出图像的隐写像素位;从目标函数层面解决正负样本的不均衡问题,提升检测精度。在基于BOSSbase v1.01数据源展开的实验中,该网络对经自适应隐写术S-UNIWARD在负载为0.4?BPP嵌入的隐写图像的像素检测准确度为0.981?74,实验验证该网络同时适用于对经非内容自适应隐写术LSB matching嵌入后的隐写图像进行隐写像素定位。     

空域序列隐写图像的密钥估计算法*

在已知嵌入消息长度的条件下,隐写密钥可以看做消息嵌入的起始位。将图像像素划分为不同类点,通过分析信息嵌入和最低位置反对各类点的影响,得到图像嵌入信息部分和未嵌入信息部分的差异,最终设计了针对空域序列LSB(least significant bits)隐写图像的密钥估计算法。实验结果表明,该算法可对隐写密钥进行快速有效的估计。     

一种新的基于双伪随机数的图像隐写算法

为提高数字图像隐写中的隐藏信息容量,提出了一种基于双伪随机数的图像隐写算法.首先介绍了伪随机数生成和双像素嵌入信息原理,然后将随机产生的整数伪随机数看作一个辅助像素值,结合载体图像中的一个像素值将两位秘密信息同时嵌入到一个载体图像像素中,从而在像素改变较小的情况下提高隐藏容量.最后分析了算法的嵌入性能,并通过仿真实验证明,该方法在提高隐藏信息容量的同时,也具有很好的安全性能.     

基于双层湿纸嵌入的自适应图像隐写算法

为了提高隐写系统抗统计检测的能力,对影响隐写算法统计安全性的因素进行了分析,得出减小载密图像统计失真的3条途径。以此为基础,提出一种基于双层湿纸嵌入的图像自适应隐写算法。该算法根据邻域相关性,选取局部环境噪声强的像素作为嵌入对象,利用“湿纸码(WPC)”在其最低有效位及次低有效位嵌入秘密信息。对未压缩图像库的实验表明,该方法比现有算法具有更好的抗统计检测性能。     

抗统计分析的BPCS隐写算法

位平面复杂度分割（BPCS）隐写在复杂度较高的位面小块嵌入秘密信息,有较好的隐蔽性,但会引起图像位面小块复杂度的统计特性的变化。通过深入分析图像位平面上相邻像素之间的相关性,本文提出了一种改进的BPCS隐写法。它更好地利用人的视觉特性,结合图像的局部特征来嵌入秘密信息。实验结果表明,改进后的算法能够抵御针对原BPCS隐写的分析方法,有较大的嵌入量,并对原图像所作的改变具有更好的不可见性。     

一种基于LSB和PVD的图像信息隐藏方法研究

为了提高图像中秘密信息的嵌入量,提出并实现了一种基于最低有效位和像素值差异的图像信息隐藏方法。通过计算两个连续像素值的差异来判断图像的一致区域和边界区域,在一致区域使用最低有效位方法,而在边界区域使用像素值差异方法。实验结果表明,结合这两种方法后信息的嵌入量是单纯使用像素值差异方法的1.59～1.97倍,并且隐写图像具有大于38 dB的峰值信噪比,保证了隐写图像的视觉质量。     

基于小波系数相关性的图像自适应空域隐写术

定义图像中像素的嵌入失真是图像自适应隐写中的关键。为提高图像自适应隐写的安全性,根据最小化嵌入失真原则,提出了一种基于小波系数相关性的图像自适应空域隐写术。首先以一维高通、低通滤波器为工具构造方向滤波器;然后沿水平、垂直、对角线方向对图像进行方向滤波,并根据小波系数与其邻域系数的相关性对失真函数进行设计;最后根据像素的嵌入失真值,利用网格码(STC)对秘密信息进行嵌入。实验结果表明,该隐写术能够将嵌入区域集中在内容复杂的纹理区域,且能够有效抵抗通用隐写检测算法的分析。     

一种针对BLOCK-DCT隐写的隐写分析改进算法

摘 要: 由于BLOCK-DCT隐写算法对平稳性质的原始图像引入了非平稳性,使得隐写图像块内部的相邻像素对与块之间的像素对的分布存在着差异。而在未嵌入消息的原始图像中它们的分布是一致的。由此利用Kolomogrov-smirnov test(k-s test)去判定图像是否隐藏信息.可是采用4邻域象素对的直接象素差的分布,在k-s test中并不怎么理想,两类概率分布之差的上界D都比较小。我们通过引入差的绝对值和8领域像素对来扩大2类象素对分布差异的方法来提高D的值,使得一些在以4邻域直接象素差分布方法中检测不出的隐写图像在改进的方法中得以检测。理论和实验都证明我们的想法是有效的。     

新的基于小波对比度的LSB密写

利用图像块的小波对比度可以判定图像各像素的LSB（Least Significant Bit）嵌入深度,以提高嵌入容量并保持载密图像的质量。提出了一种新的基于小波对比度的密写方案,将图像块像素的低4位置0后进行小波变换并计算小波对比度作为嵌入深度的判决信息,由嵌入深度逐块逐像素密写。实验结果表明,该方法不仅保持了较高的嵌入容量,而且提取秘密信息时不需任何附加信息。     

一种调色板聚类的信息隐藏算法*

针对目前调色板信息隐藏方法的弱点, 结合聚类分析技术, 提出了一种新的基于调色板聚类的信息隐藏算法ClusterStego 。根据定义的相似性测度和聚类准则函数, 通过对由调色板形成的RGB 空间进行合理聚类来隐藏信息, 具有良好的直观效果。同时由于引入了质量系数的概念, 使得该算法具有隐藏质量可控性, 而且其隐藏容量远大于EzStego 等算法, 可达到每像素87. 5% 。     

Reversible watermarking based on multiple prediction modes and adaptive watermark embedding

A new reversible watermark scheme based on multiple prediction modes and adaptive watermark embedding is presented. Six prediction modes fully exploiting strong correlation between any pixel and its surrounding pixels, are designed in this paper. Under any prediction mode, each to-be-predicted pixel must be surrounded by several pixels (they constitute a local neighborhood, and any modification to this neighborhood is not allowed in the embedding process). This neighborhood has three main applications. The first one is that when it is exploited to interpolate some to-be-predicted pixel, the noticeable improvement in prediction accuracy is obtained. The second one is that its variance is employed to determine which classification (i.e., smooth or complex set) its surrounded pixel belongs to. For any to-be-predicted pixel, the number of embedded bits is adaptively determined according to this pixel’s belonging. The last one is that we can accurately evaluate the classification of watermarked pixels by analyzing the local complexity of their corresponding neighborhoods on the decoding side. Therefore, the payload can be largely increased as each to-be-predicted pixel in the smooth set can possibly carry more than 1 bit. Meanwhile, the embedding distortion is greatly controlled by embedding more bits into pixels belonging to smooth set and fewer bits into the others in complex set. Experimental results reveal the proposed method is effective.     

High-capacity image hiding scheme based on vector quantization

A novel grayscale image hiding scheme that is capable of hiding multiple secret images into a host image of the same size is proposed in this paper. The secret images to be hidden are first compressed by vector quantization with additional index compression process. Then, the compressed secret images are encrypted and embedded into the least-significant bits of the host pixels. To provide good image quality of the stego-image, the modulus function and the image property are employed to hide the secret bits into the host pixels and determine the number of hidden bits in each host pixel, respectively. According to the results, the proposed scheme provides a higher hiding capacity and a higher degree of security than that of the virtual image cryptosystem.     

An LSB based novel data hiding method using extended LBP

In this paper we have proposed an LSB based data hiding method. Here, first cover image is encoded by LBP based method (call as ELBP). In ELBP method, a 3 × 3 block is encoded by k bits (1 ≤ k ≤ 9) respect to the central pixel. For data hiding purpose, the cover image is encoded by ELBP then r (1 ≤ r < k) bits are embedded into each neighbor pixel by replacing least r bits of the encoded stream and then modified pixel is processed by OPAP method to improve the quality of the stego image. Proposed method gives high quality stego images with higher embedding rate compare to the state-of-the-art methods. The security of the proposed data hiding method is tested against the attacks like RS steganalysis and Chi-square attack and these methods fail to detect hidden data.

     

A novel data hiding for color images based on pixel value difference and modulus function

This paper proposes a novel data hiding method using pixel-value difference and modulus function for color image with the large embedding capacity(hiding 810757 bits in a 512×512 host image at least) and a high-visual-quality of the cover image. The proposed method has fully taken into account the correlation of the R, G and B plane of a color image. The amount of information embedded the R plane and the B plane determined by the difference of the corresponding pixel value between the G plane and the median of G pixel value in each pixel block. Furthermore, two sophisticated pixel value adjustment processes are provided to maintain the division consistency and to solve underflow and overflow problems. The most importance is that the secret data are completely extracted through the mathematical theoretical proof.     

An image authentication method by applying Hamming code on rearranged bits

In 2007, Chan and Chang proposed an image authentication method using the Hamming code technique. The parity check bits were produced from pixels by using the Hamming code technique, and the produced bits were embedded in other pixels. When recovering, the method had to predict the value of the most-significant bit of each tampered pixel first. Then, the tampered pixel was able to be recovered by referring to the predicted bit and its parity check bits. However, using the most-significant bit is unsuitable because of the risk of making an incorrect prediction. In this paper, the parity check bits are produced from pixels whose bits have been rearranged. This way, the value of the most-significant bit of each tampered pixel can be determined according to its parity check bits. The recovery procedure is also modified to accommodate the rearranging procedure. The experimental results show that the proposed method has a better ability to recover the tampered areas, compared with Chan and Chang’s method. Moreover, the quality of the authenticated images of the proposed method is also higher than that of Chan and Chang’s method.     

An adaptive steganographic method based on the measurement of just noticeable distortion profile

This paper presents an adaptive steganographic method based on just noticeable distortion (JND) profile measurement. According to the input requirements, our method can produce a higher quality or higher embedding capacity stego-image. In the embedding secret data bits into a target pixel process, four different impact factors are utilized to compute how much information can be embedded and what the stego-pixel value will be. These are difference values that represent the correlation between neighboring pixels, the JND value of the target pixel, a predefined embedding capacity control factor, and the contents of various lengths secret data bits. The proposed method embeds more secret data bits within complex areas and less data bits in smooth areas. The difference between the target pixel and the stego-pixel values is controlled, as far as possible, to less than or equal to the JND value of the target pixel. Thus, the stego-image maintains good imperceptible property. In the extraction phase, the embedded secret data can be extracted from the stego-image without referencing the original image and the JND profile. The experimental results show that our method improves stego-image quality and conspicuously increases the embedding capacity at the same time.     

A high-capacity reversible data hiding scheme based on sorting and prediction in digital images

In this paper, a new reversible data hiding method based on sorting and prediction in digital images is proposed. The proposed method can embed two bits in the 3 × 1 sub-block at maximum by dividing into two groups - min and max groups. The pixel pairs of the min group and max group are firstly predicted and then are modified to embed the secret bits. The reversibility is guaranteed since the order of pixel pairs of the sub-block is not changed after embedding a secret bit into two groups. The experimental results show that the proposed method provides high embedding capacity than the previous works.     

High-capacity PVO-based reversible data hiding scheme using changeable step size

Reversible data hiding (RDH) algorithm based on pixel-value-ordering (PVO) has received widespread attention because of its excellent performance. PVO algorithm divides the host image into non-overlapped equal-sized blocks, then achieves data embedding by modifying the maximum and minimum values of each block. Every pixel block can be a host of watermark data, so the smaller number of pixel blocks limit the embedding capacity (EC). In our work, a novel PVO with changeable step size (CPVO) is presented which can choose suitable step size based on the number of watermark data bits, even that allow one block to overlap other ones. Take the block size 2 × 2 as an example, we can set step size 2 × 1 or another one in CPVO. Consequently, with a block selection skill based on the noise level (NL) of a pixel block, CPVO can embed more hidden data bits into a host image. Compared to the original PVO-based schemes, experimental results show that our proposed scheme increases the EC by 2 ～ 3 times, and the marked image quality keeps much higher, even outperforms some other state-of-the-art works in some test images.